ORIG_HEAD

Is there a really good reason why git-pull-script runs are deleting
that file now?

All my scripts use ORIG_HEAD so that I can look at the changelog after
I pull from someone.  There is now no record left around of what the
head was before the pull.

I could take the SHA1 output by the git-pull-script, but it'd be nice
if it was sitting under .git somewhere in case I forget to record
that information somewhere.

This is my "git-mklog" script, for example:

#/bin/sh

git-rev-tree HEAD ^ORIG_HEAD | sort -n | cut -d' ' -f2 >commit-list
cat commit-list | git-diff-tree --stdin -p | diffstat -p1 
echo ''
for i in $(cat commit-list)
do
    git-diff-tree -s -v ${i}
    git-diff-tree -p ${i} | diffstat -p1
    echo ''
done

rm -f commit-list

On Mon, 20 Jun 2005, David S. Miller wrote:

Is there a really good reason why git-pull-script runs are deleting
that file now?

No. I've cleaned it up a bit, and codified the stuff we leave around.

I also changed "git fetch" to write FETCH_HEAD instead of MERGE_HEAD, 
because that's obviously what it is (it's perfectly fine to fetch things 
for other reasons, like just checking what somebody else has, and you 
might not ever intend to merge it anyway).

		Linus